Assessment of a new dynamic RANS approach for the study of parachute-capsule aerodynamics

Nowadays, the aerodynamic study of bluff bodies plays a crucial role in industrial applications. These geometries are characterized by highly unsteady flow physics, dominated by vortex structures distributed across a wide range of scales, which dictate the resulting loads and overall flow behavior. A major challenge in simulating such cases lies in achieving an optimal trade-off between computational cost and solution accuracy. This difficulty arises because conventional solvers sit at opposite extremes: traditional Reynolds-Averaged Navier-Stokes (RANS) models offer low computational cost, but they yield poor accuracy in highly unsteady regions, since the entire turbulent spectrum is modeled rather than resolved. On the other hand, Large Eddy Simulation (LES) provides high-fidelity results, but its computational cost is often unsustainable for practical applications. In this context, the dynamic RANS model represents a promising compromise by allowing turbulence resolution to adapt to the local grid size, resolving turbulent structures up to the limit imposed by the mesh. This thesis focuses on testing how well the dynamic RANS model performs when applied to the complex aerodynamics of a parachute-capsule system. To verify its accuracy, the numerical predictions are benchmarked against a reference LES solution obtained on a much finer grid, assessing how the method behaves when mesh refinement is restricted to the critical regions of the flow. The comparison demonstrates that the dynamic RANS model delivers good agreement with the LES results despite relying on a significantly coarser grid, offering a much more efficient alternative to resolve the aerodynamics of bluff bodies in industrial applications.